ARTICLE

A Detailed Study of Irradiation Requirements, Towards an Efficient

Photochemical Wohl-Ziegler Protocol in Flow

Holly E. Bonfield,[a] Jason D. Williams,[a][b] Wei Xiang Ooi,[a] Stuart G. Leach,[a] William J. Kerr[b] and Lee

J. Edwards*[a]

Dedicated to the memory of Dr. Matthew P. John

Abstract: A platform has been developed to enable standardization

of light sources, allowing consistent scale-up from high throughput

screening in batch to flow, using the same pseudo-monochromatic

light source. The impact of wavelength and light intensity on a

photochemical reaction was evaluated within this platform, using the

Wohl-Ziegler benzylic bromination of 4-methyl-3-

(trifluoromethyl)benzonitrile with N-bromosuccinimide as a model

system. It was found that only 40% of the maximum light intensity was

required whilst still maintaining reaction rate, allowing more reliable

temperature control and lower energy consumption. The optimized

reaction conditions were subsequently applied to a range of

synthetically relevant (hetero)aromatic compounds under continuous

conditions, exploring the scope of the process within a mild and

scalable protocol.

Introduction

In recent years, the use of synthetic organic photochemistry

has increased significantly. Within an industrial setting,

photochemical methods have historically employed UV radiation

to directly excite organic molecules.[1] However, current research

endeavors are focusing on harnessing visible light for synthetic

applications, and these have started to become of interest

industrially as a viable route to pharmaceutically active

compounds.[2] With an increasing number of domestic light

sources available (LED/CFL technologies), an array of synthetic

methodologies utilizing visible light have been published and

implemented in a variety of applications, with improved selectivity

credited to the lower energy wavelengths used.[3] However,

despite substantial advances, light sources used by synthetic

organic chemists in photochemical transformations still remain

poorly understood or not considered at all, with reaction

conditions optimized to their light source at a single power setting.

Internally it has been found that moving to a new light source on

up-scaling often necessitates subsequent re-optimization.

The nature of the light source is central to the understanding

of photochemical reactions. Light is an essential component in

these transformations, yet sources used in synthetic applications

are often poorly characterized broad-spectrum lamps.[4] Although

this seemingly allows a convenient “one size fits all” solution, a

more targeted approach is now possible with the development of

pseudo-monochromatic light sources. This targeted approach

allows for the in-depth development of a robust and fully optimized

reaction protocol, particularly for the delivery of active

pharmaceutical ingredients (APIs). Modern high power pseudo-

monochromatic LEDs offer significant benefits over other

commonly used industrial light sources, not only in their precise

spectral output but also in their comparatively low operating

temperature, which is a result of enhanced efficiency.[5] This

delivers a marked improvement upon medium-pressure mercury

lamps, for example, which have multiple discrete emittance bands

between 200-500 nm, leading to unwanted photochemical

reactions. In addition, these lamps typically operate between 600-

800 °C,[6] meaning that temperature control is a significant

challenge, with unwanted thermal side reactions often observed.

The use of narrow band filters can restrict spectral output to the

desired wavelength, but this can significantly reduce the

transmittance, making the approach wasteful, particularly when

considering the cooling required to remove excess heat from the

lamp and/or the reaction.

Since its first report almost a century ago, the Wohl-Ziegler

benzylic bromination has been regularly implemented in

synthesis.[7] The majority of applications use radical initiators,

such as peroxides[8] or azo compounds,[9] to mediate the

homolysis of bromine.[10] However, light is a cheap and renewable

alternative to traditional initiation approaches, delivering a safer

and more sustainable transformation. Consequently, the

photochemical Wohl-Ziegler reaction is well-precedented, with

recent studies using a range of light sources.[11] Our experience

with this reaction using medium-pressure mercury lamps, and the

evaluation of batch-to-batch variability within domestic light

sources, demonstrated a need for improved understanding of the

light source. This has prompted us to undertake a detailed

investigation into the effect of the character of the light source on

the Wohl-Ziegler bromination. To our knowledge, no such study

has previously been conducted, and this has resulted in the use

of a wide range of conditions, including the unnecessary use of

UV irradiation, or a combination of both a radical initiator and

irradiation.[12]

Photochemical benzylic brominations have previously been

demonstrated under continuous conditions,[13-14] indicating the

[a] H. E. Bonfield, W. X. Ooi, L. J. Edwards, Dr. J. D. Williams, Dr. S. G.

Leach

API Chemistry, GlaxoSmithKline Medicines Research Centre,

Gunnels Wood Road, Stevenage, Hertfordshire SG1 2NY (UK)

E-mail: lee.j.edwards@gsk.com

[b] Prof. Dr. W. J. Kerr, Dr. J. D. Williams

Department of Pure and Applied Chemistry

WestCHEM, University of Strathclyde

295 Cathedral Street, Glasgow, Scotland, G1 1XL (UK)

Email: w.kerr@strath.ac.uk

Supporting information for this article is given via a link at the end of

the document.

ARTICLE

potential scalability of this protocol, and thus its industrial

applicability.[15] When considering the development of an

industrial process, all reaction variables must be fully understood

to ensure consistent product yield and quality, and in a

photochemical reaction, irradiation character is no exception. In

this respect, the light source should be optimized to the required

wavelength and intensity, to minimize energy consumption and

by-product formation.[16] Tuning the output of a light source can

bestow further advantages, particularly by reducing the cooling

requirement of the reaction system, which can present a

challenge on larger scales.

Results and Discussion

As a model for our initial studies into the photochemical

Wohl-Ziegler reaction, we selected 4-methyl-3-

(trifluoromethyl)benzonitrile 1, a challenging, electron-poor

benzylic bromination substrate, which has been shown to require

prolonged heating when using a chemically initiated protocol

(Scheme 1).[17] Moreover, this substrate is of pharmaceutical

interest making it an appropriate example to study in detail,

enabling the factors which affect the reaction selectivity and

productivity to be established.

Scheme 1. The benzylic bromination of 1, studied in this report, under

photochemical conditions.

In order to select the most suitable wavelength for irradiation,

the reaction components were analyzed by UV-visible absorption

spectroscopy in order to obtain confirmation of the photoactive

species (Figure 1). Unsurprisingly, this revealed no absorbance

above 350 nm for the reaction substrate 1, N-bromosuccinimide

(NBS), or a mixture of the two. However, the absorption spectrum

of a sample of commercially obtained NBS, without prior

purification, showed significant overlap with that of bromine (λmax

= 395 nm), indicating that low levels of bromine are responsible

for the photochemical initiation of this benzylic bromination. The

low Br-Br bond strength (190 kJ mol−1)[18] suggests that light with

Figure 1. UV-visible absorption spectroscopy (300-750 nm) of the reaction

components in the benzylic bromination of 4-methyl-3-

(trifluoromethyl)benzonitrile (1).

a wavelength shorter than 630 nm is sufficiently energetic to

promote homolysis. Accordingly, an LED which has a spectral

output close to 395 nm should be most suitable in this reaction,

leading to the use of visible light LEDs (λ > ~400 nm). Following

on from these observations, it was realized that the quantity of

bromine present in a sample of NBS can be determined in a

straightforward manner using UV-visible absorption spectroscopy.

Varying levels of bromine in NBS could impact the reaction rate

and impurity profile, therefore an assay method to determine the

bromine levels is desirable and, to our knowledge, has not been

discussed in prior publications. Construction of an absorption

calibration curve, using known concentrations of bromine, allowed

the quantity of bromine in a commercial sample of NBS to be

quantified as 3.1 mol%. This dropped to an undetectable level (<

0.1 mol%) following recrystallization from water, indicating that

the remaining trace bromine is sufficient to initiate the reaction.

With this information in hand, investigation of reaction

conditions was performed using high throughput screening (48-

well plates), under irradiation by pseudo-monochromatic LEDs

(Table 1). Initial screening reactions at a range of substrate and

NBS concentrations in acetonitrile gave promising results, with up

to 78% monobrominated product formation in one hour (entries 1-

5). However, alongside the expected benzylic bromide product 2,

an appreciable level of the corresponding dibrominated

compound, 4-(dibromomethyl)-3-(trifluoromethyl)benzonitrile 3

was observed, presumably due to the poor energetic

differentiation between H-atom abstraction from the starting

material 1 and the desired product 2.[19] Furthermore, this second

bromination led to consumption of additional NBS, such that

unreacted starting material 1 also remained. It is hypothesized

that bromine formation from NBS is favoured under Brønsted

acidic conditions,[20] allowing the concentration of bromine radical

sufficient for reaction to be reached more quickly. The addition of

acetic acid (entry 6 versus entry 8) enhanced the rate of reaction

significantly, allowing near-complete conversion in just five

minutes, while extending the reaction time to 60 min under these

conditions offering no significant benefit (entry 7). Increasing the

number of NBS equivalents to 1.5 (entry 8 versus entry 9) showed

similar reaction composition at five minutes but also

0.0

0.5

1.0

1.5

2.0

0.00

0.02

0.04

0.06

0.08

300 400 500 600 700

Bro

min

e ab

sorb

ance

/ A

U

Ab

sorb

ance

/ A

U

Wavelength / nmCommercial NBS Recrystallized NBS Substrate Bromine

ARTICLE

overbromination to by-product 3 over time (entry 9 versus entry

10). Investigation of acetic acid loading (entries 11 to 14) indicated

that either increasing or decreasing the quantity of acetic acid

from a 1:1 ratio with MeCN causes deviation from the optimal yield

of product 2. When varying the wavelength of the light source (405

nm in entries 1-14, 420 nm in entry 15, and 450 nm in entry 16),

a similar final reaction composition was observed. However, upon

more detailed examination, the 405 nm light source was found to

mediate the transformation at a faster rate, so its use was

continued throughout the remainder of this investigation.[21]

Additional control reactions were performed, which

confirmed that no reaction occurred in the absence of light, or in

the presence of a radical inhibitor species.[22] Importantly, it was

also observed that continual irradiation of the reaction mixture

was required, as no further reaction was observed following

removal of the light source, prior to reaction completion. This

implies that radical chain propagation processes alone are not

sufficient to sustain the reaction, and continued energy input from

an external source is necessary.

Further understanding of the course of reaction was sought

through more in-depth reaction profiling. This was used to

compare the rate of starting material 1 consumption (Figure 2a),

alongside formation of product 2 and by-product 3, under

irradiation at different light intensities. These varied intensities

were accessed using tunable 405 nm LEDs, whose input power

was found to have a linear relationship with output intensity, as

measured in lux (Figure 2b). As light intensity was decreased, a

more pronounced initiation period was observed prior to reaction

progression. This was most notable in the reaction run using 30

mA current (giving the lowest possible output power), which

appeared to have an initiation period of approximately 2 minutes.

Initiation periods have been studied previously in halogenation

reactions which use a radical initiator [23] but, to our knowledge,

never within the photochemically-initiated variant. Substantially

shorter initiation periods were exhibited at higher light intensities,

until no significant rate difference was observed, between 200 mA,

300 mA and 400 mA currents [24]. This implies that the reaction

mixture is light-saturated at these higher light intensities, and the

reaction is therefore operating within an entirely reagent-limited

Table 1. Reaction optimization: varying concentration, NBS loading and solvent composition.[a]

Entry Solvent Time (min) Conc. (M) NBS loading (eq.) Remaining 1 (%)[b] 2 (%)[b] 3 (%)[b]

1 MeCN 60 0.1 1.05 12 66 1

2 MeCN 60 0.2 1.05 14 70 2

3 MeCN 60 0.3 1.05 11 74 2

4 MeCN 60 0.4 1.05 13 74 3

5 MeCN 60 0.5 1.05 9 78 4

6 MeCN 5 0.3 1.05 80 17 0

7 MeCN/AcOH (1:1) 60 0.3 1.05 5 78 5

8 MeCN/AcOH (1:1) 5 0.3 1.05 12 77 4

9 MeCN/AcOH (1:1) 5 0.3 1.5 7 80 7

10 MeCN/AcOH (1:1) 60 0.3 1.5 0 54 29

11 MeCN/AcOH (1:1) 5 0.1 1.5 5 75 5

12 MeCN/AcOH (99:1) 5 0.1 1.5 81 17 0

13 MeCN/AcOH (8:2) 5 0.1 1.5 42 52 1

14 AcOH 5 0.1 1.5 2 70 19

15[c] MeCN/AcOH (1:1) 5 0.1 1.5 5 79 8

16[d] MeCN/AcOH (1:1) 5 0.1 1.5 9 76 4

[a] Light source was set to 200 mA current. [b] Quoted yields are HPLC %area. [c] Reaction performed using 420 nm LEDs set to 200 mA current. [d] Reaction

performed using 450 nm LEDs set to 200 mA current.

ARTICLE

kinetic regime. This can be explained by the relatively low

concentration of bromine present in these reactions, which allows

substantially higher light transmission than in many other

photochemical reactions. 200 mA current was chosen for all

further reactions, based on this observation.

The influence of bromine content on the length of initiation

period was demonstrated by comparison of time courses between

recrystallized and commercial NBS (Figure 3). Upon exclusion of

bromine, a more distinct initiation period was observed, with a

peak quantity of product 2 observed after 5.5 minutes, compared

to 4.5 minutes. However, both reactions displayed the same

reaction composition at peak conversion, with 3-5% remaining

starting material 1, 77-78% desired product 2 and 8% of the

unwanted overbrominated by-product 3. Since the commercial

NBS reached peak conversion slightly faster, it was used in all

other reactions.

The optimal reaction conditions (using 405 nm LEDs, at 200

mA current (40% of the maximum light intensity)) were applied to

the reaction in flow, where the residence time was optimized to

achieve maximum conversion to monobrominated product 2,

whilst minimizing the time allowed for formation of dibrominated

by-product 3. From the 5 minute reaction time required in batch

(Table 1, entry 9), the optimal residence time was decreased to

2.5 minutes (Figure 4a), due to the improved light penetration and

mixing achieved when using a highly turbulent microreactor (4.09

mL volume) in flow.[25-26] Upon reducing the number of NBS

equivalents to 1.05, a longer residence time of 5 minutes (Figure

4b) was required to maintain an excellent yield of the desired

product 2. Under these conditions, one gram of material was

processed, with a consistent reaction profile over the 20 minute

run duration, leading to a 69%

Figure 4. Product distribution of benzylic bromination of 1 in flow, using different

residence times, to determine the optimal flow rate. λ = 405 nm, current = 200

mA. a) Flow rate optimization using 1.5 eq. NBS, displaying an optimal product

2 yield at 2.5 minute residence time. b) Flow rate optimization using 1.05 eq.

NBS, displaying an optimal product 2 yield at 5 minute residence time.

0

20

40

60

80

100

0 2 4 6 8

Rem

ain

ing

1/

%

Reaction time / min

30 mA 100 mA 200 mA 300 mA 400 mA

R² = 0.9991

0

5000

10000

15000

20000

25000

0 100 200 300 400 500 600

Lum

ino

us

flu

x /

lx

Input current / mA

b)

0

20

40

60

80

100

1 1.5 2 2.5 3.5 4.5 5

Yiel

d /

%

Residence time / min

3

1

2

a)

a)

0

20

40

60

80

100

4.5 5 5.5 6

Yiel

d /

%

Residence time / min

3

1

2

0

20

40

60

80

100

0 2 4 6 8 10

Yiel

d 2

/ %

Reaction time / min

ARTICLE

Table 2. Substrate scope examined under flow conditions. λ = 405 nm, 200 mA current. 1 mmol of substrate was processed.

[a] Conversion based on percentage of starting material consumed by HPLC %area. [b] Yield of the desired product, as determined by HPLC %area. [c] Isolated

yield of the desired product is shown in parentheses.

isolated yield of the monobrominated product 2.

The procedure was then applied to the benzylic bromination

of a range of electron-deficient aromatic and heteroaromatic

compounds (Table 2). NBS equivalents and residence times for

each substrate were selected based on an initial substrate screen, [28] and no further optimization was performed. In flow all

substrates underwent facile conversion to their corresponding

benzylic bromides within very short residence times. The

presence of a boronic ester (entry 1) was well tolerated, providing

the desired product 4 in a residence time of just 2 minutes.

Selective formation of product 5 was achieved (entry 2), and no

α-bromination of the ketone was observed. This has been known

to occur in the presence of elemental bromine under acidic

conditions.[29] A heavily functionalized pyrimidine core was

brominated on its ethyl group (entry 3) in just 1.36 minutes,

whereas two other heterocyclic cores (entries 4 and 5) required

extended residence times (18 and 9 minutes, respectively) and a

higher NBS loading, in order to achieve good conversion. Ester-

containing aromatics (entries 6 and 7) were brominated smoothly,

yet those furnished with a nitro group (entries 8 and 9) required

Entry Product Compound NBS loading

(eq.)

Residence time

(min)

Conversion

(%)[a] Product yield

(%)[b][c]

1

4 1.05 2.0 100 94 (80)

2

5 1.05 1.36 95 83 (53)

3

6 1.05 1.36 97 87 (35)

4

7 1.5 18 82 57 (43)

5

8 1.5 9 87 69 (49)

6

9 1.05 1.36 98 91 (89)

7

10 1.05 1.36 94 85 (30)

8

11 1.5 4.54 96 74 (53)

9

12 1.5 1.36 91 80 (47)

ARTICLE

an increased NBS loading. Notably, excellent temperature control

was maintained throughout these reactions, typically keeping the

flow plate temperature below 30 °C, with the exception of the

examples using slower flow rates (entry 4 reached a maximum

temperature of 60 °C and entry 5 reached 40 °C).[30] It is possible

that in these examples, the higher temperature may have resulted

in poorer selectivity for the desired monobrominated products.

The flow protocol was examined on a larger scale, by

running the reaction continuously for 5.5 hours (Figure 5a). Due

to the fast reaction rates of the protocol, this elongated run

encompassed 243 residence times, within which 46.6 g of starting

material was processed. The reaction composition (Figure 5b)

was monitored throughout, displaying a steady profile after the

initial equilibration period. During this period the reaction temperature was also measured at the reactor entrance and exit,

and within the reactor itself, in order to accurately illustrate the temperature profile (Figure 5c).[31] The reactor temperature (T3)

was found to reach steady state between 40-42 °C, and

maintained this throughout the run. This larger scale reaction

demonstrates the industrial applicability of the developed

procedure, by exemplifying consistent reaction profile and

temperature control within a continuous flow process. The desired

product 10 was then isolated via an unoptimized recrystallization

in 40% yield (27.5 g), comparable to the smaller scale reaction

(Table 2, entry 7). It was subsequently found that 10 could be

generated in a solution yield of 58% at 400 mA current, using a

significantly shorter 0.20 minute residence time. This compares

favourably with the 57% solution yield achieved using 200 mA

current with a 1.36 minute residence time, as throughput is

increased 6-fold.[32]

Finally, our optimization methodology was applied to 15

(Table 2, entry 6) to identify the optimal intensity and flow rate for

the generation of 9 in higher throughput. Different light intensities

and flow rates were screened (Figure 6b). It was found that 350

mA current allowed a flow rate of 10 mL/min (0.41 minute

residence time). These conditions were advantageous when

compared to those from the original optimization (200 mA current,

3 mL/min flow rate 1.36 minute residence time), and thus were

used to process 30 g of 15. An additional 5 g of 15 were processed

using the lower light intensity and longer residence time. The

reaction profiles for both sets of conditions were similar by HPLC[33] and isolated yields of 62% (25 g, with mother liquors

containing a further 2.5% (1.07 g)) and 71% (4.8 g, with mother

liquors containing a further 18% (1.22 g)) were achieved

respectively, via an unoptimized recrystallization. These yields

are consistent with the solution yields realized in Figure 6b.

Conclusions Our studies have shown that pseudo-monochromatic visible

light LEDs provide effective irradiation for the photochemical

0

10

20

30

40

50

60

70

80

90

100

0 100 200 300 400

Yiel

d /

%

Reaction time / min

13

10

14

15

20

25

30

35

40

45

0 50 100 150 200 250 300 350

Tem

per

atu

re /

°C

Processing time / min

T1

T2

T3

a)

c)

b)

a)

Figure 6. a) Reaction scheme for the large scale bromination of ester 15, to

form its benzylic bromide 9. b) Chart showing the solution yield of reaction at

different flow rates and light intensities.

Figure 5. a) Reaction scheme for the large scale bromination of ester 13, to

form its benzylic bromide, 10. b) Chart showing the reaction profile throughout

the 5.5 hour operating time. Reaction yields are measured by HPLC,

using %Area values. c) Reaction temperature profile throughout the 5.5 hour

operating time. T1 = input temperature, T2 = output temperature, T3 = reactor

temperature. 0

20

40

60

80

100

0 100 200 300 400 500

Solu

tio

n Y

ield

9 /

%

Current / mA

3 mL/min 5 mL/min 7.5 mL/min 10 mL/min

b)

ARTICLE

Wohl-Ziegler benzylic bromination using NBS. By using a platform

capable of varying light intensity, it has been possible to develop

a protocol that uses only a fraction of the maximum LED power,

thereby developing our understanding of power and cooling

requirements, when scaling photochemical processes from the

lab to an industrial platform.

Scaling up of this process was demonstrated from

screening through to running the reaction continuously for 5.5

hours in flow with the same pseudo-monochromatic light source.

Consequently, reliable temperature control has been achieved in

flow, for a reproducible reaction, with minimal energy wastage.

We have also applied this methodology to a range of other

substrates, where excellent conversions were achieved with a

residence time of less than 2 minutes in many cases.

We have demonstrated how our strategy of optimizing the

light source to suit the required chemical transformation can be

applied. In doing so we have developed generic conditions for the

exemplar substrate that can be applied to other substrates and

methodology that can be applied to further optimize each

substrates, either for quality or throughput. We anticipate that an

improved emphasis on the nature and power of light source used,

particularly by synthetic organic chemists, will allow development

of more efficient and selective photochemical processes which

are directly transferrable to industry.

Experimental Section

Detailed experimental procedures for all experiments, and characterization

data for all products are available in the supporting information.

General batch protocol: 1 mL of a 0.1 M solution containing 4-methyl-3-

(trifluoromethyl)benzonitrile (1 eq.) and N-bromosuccinimide (1.5 eq.) in

1:1 MeCN:AcOH was charged to 48 × 2 mL vials and each inserted into a

48-well plate on the photoreactor. The vials were irradiated at a chosen

wavelength and light intensity, for the desired length of time. Time course

experiments were performed by removing a reaction vial every 30 seconds,

for HPLC analysis.

General flow protocol: A 0.3 M solution of bromination substrate and N-

bromosuccinimide (1.05 - 1.5 eq.) in 1:1 MeCN:AcOH was irradiated with

an array of 48 × LEDs (405 nm wavelength, set to 200 mA input current),

whilst being passed through a glass plate flow reactor at an optimized flow

rate. After the required volume of reaction mixture had entered the reactor,

the reaction mixture was collected for a time equivalent to four reactor

volumes. The desired product was isolated as specified in each individual

example.

Acknowledgements

The authors thank Mr Peter Gray and Mr Khushaal Sharma for

their aid in developing the photoreactor, Dr. David M. Lindsay for

assistance in manuscript preparation, Mr. Thomas Atherton for

mass spectrometry support and Mr Graham Rutherford for HiTec

Zang programming support. J.D.W. would also like to thank

GlaxoSmithKline for financial support.

Keywords: Flow Chemistry • Halogenation • Microreactors •

Photochemistry • Reaction mechanisms

[1] Reviews of direct excitation UV photochemistry in organic synthesis: a)

N. Hoffmann, Chem. Rev. 2008, 108, 1052–1103; b) B. D. A. Hook, W.

Dohle, P. R. Hirst, M. Pickworth, M. B. Berry, K. I. Booker-Milburn, J. Org.

Chem. 2005, 70, 7558–7564; c) J. D. Winkler, C. M. Bowen, F. Liotta,

Chem. Rev. 1995, 95, 2003–2020; d) M. T. Crimmins, Chem. Rev. 1988,

88, 1453–1473.

[2] Selected examples of photochemical processes in drug discovery and

development: a) J. J. Douglas, M. J. Sevrin, C. R. J. Stephenson, Org.

Process Res. Dev. 2016, 20, 1134–1147; b) M. Fischer, Angew. Chem.

1978, 90, 17–27; M. Fischer, Angew. Chem. Int. Ed. 1978, 17, 16–26; c)

D. Cambié, C. Bottecchia, N. J. W. Straathof, V. Hessel, T. Noël, Chem.

Rev. 2016, 116, 10276–10341; d) Z. Amara, J. F. B. Bellamy, R. Horvath,

S. J. Miller, A. Beeby, A. Burgard, K. Rossen, M. Poliakoff, M. W. George,

Nat. Chem. 2015, 7, 489–495; e) J. Turconi, F. Griolet, R. Guevel, G.

Oddon, R. Villa, A. Geatti, M. Hvala, K. Rossen, R. Göller, A. Burgard,

Org. Process Res. Dev. 2014, 18, 417–422.

[3] Reviews of visible light photochemistry in organic synthesis: a) C. K. Prier,

D. A. Rankic, D. W. C. Macmillan, Chem. Rev. 2013, 113, 5322–5363;

b) Y. Xi, H. Yi, A. Lei, Org. Biomol. Chem. 2013, 11, 2387–2403; c) J.-R.

Chen, X.-Q. Hu, L.-Q. Lu, W.-J. Xiao, Chem. Soc. Rev. 2016, 45, 2044–

2056; d) M. H. Shaw, J. Twilton, D. W. C. MacMillan, J. Org. Chem. 2016,

81, 6898–6926.

[4] Selected examples of synthetic chemistry using broad spectrum light

sources: a) A. Yavorskyy, O. Shvydkiv, N. Hoffmann, K. Nolan, M.

Oelgemöller, Org. Lett. 2012, 14, 4342–4345; b) Y. S. M. Vaske, M. E.

Mahoney, J. P. Konopelski, D. L. Rogow, W. J. McDonald, J. Am. Chem.

Soc. 2010, 132, 11379–11385; c) D. C. Harrowven, M. Mohamed, T. P.

Gonçalves, R. J. Whitby, D. Bolien, H. F. Sneddon, Angew. Chem. 2012,

124, 4481–448; D. C. Harrowven, M. Mohamed, T. P. Gonçalves, R. J.

Whitby, D. Bolien, H. F. Sneddon, Angew. Chem. Int. Ed. 2012, 51,

4405–4408; d) J. Liu, Q. Liu, H. Yi, C. Qin, R. Bai, X. Qi, Y. Lan, A. Lei,

Angew. Chem. 2014, 126, 512–516; J. Liu, Q. Liu, H. Yi, C. Qin, R. Bai,

X. Qi, Y. Lan, A. Lei, Angew. Chem. Int. Ed. 2014, 53, 502–506; e) L. J.

Allen, P. J. Cabrera, M. Lee, M. S. Sanford, J. Am. Chem. Soc. 2014,

136, 5607–5610.

[5] a) C. Le, M. K. Wismer, Z.-C. Shi, R. Zhang, D. V. Conway, G. Li, P.

Vachal, I. W. Davies, D. W. C. MacMillan, ACS Cent. Sci. 2017, 3, 647–

653; b) Y. Su, N. J. W. Straathof, V. Hessel, T. Noël, Chem. Eur. J. 2014,

20, 10562–10589; c) D. H. Chen, X. Ye, K. Li, Chem. Eng. Technol. 2005,

28, 95–97.

[6] J. P. Knowles, L. D. Elliott, K. I. Booker-Milburn, Beilstein J. Org. Chem.

2012, 8, 2025–2052.

[7] Seminal publication, and reviews of the Wohl-Ziegler reaction: a) A. Wohl,

Berichte der Deutschen Chemischen Gesellschaft 1919, 52, 51–63; b) C.

Djerassi, Chem. Rev. 1948, 43, 271–317; c) I. Saikia, A. J. Borah, P.

Phukan, Chem. Rev. 2016, 116, 6837–7042.

[8] Selected recent examples: a) J. Emmerich, C. J. Van Koppen, J. L.

Burkhart, Q. Hu, L. Siebenbürger, C. Boerger, C. Scheuer, M. W.

Laschke, M. D. Menger, R. W. Hartmann, J. Med. Chem. 2017, 60,

5086–5098; b) C. J. Evoniuk, G. D. P. Gomes, S. P. Hill, S. Fujita, K.

Hanson, I. V. Alabugin, J. Am. Chem. Soc. 2017, 139, 16210–16221; c)

C. Chen, X. Chen, X. Zhang, S. Wang, W. Rao, P. W. H. Chan, Adv.

Synth. Catal. 2017, 359, 4359–4368.

[9] Selected recent examples: a) D. Feng, J. Wan, F. Teng, X. Ma, Catal.

Commun. 2017, 100, 127–133; b) W. Xun, B. Xu, B. Chen, S. Meng, A.

S. C. Chan, F. G. Qiu, J. Zhao, Org. Lett. 2018, 20, 590–593; c) H.

Shimojo, K. Moriyama, H. Togo, Synthesis 2015, 47, 1280–1290; d) M.

Bersellini, G. Roelfes, Org. Biomol. Chem. 2017, 15, 3069–3073; e) S.

Lee, O. S. Kwon, C. S. Lee, M. Won, H. S. Ban, C. S. Ra, Bioorg. Med.

Chem. Lett. 2017, 27, 3026–3029; f) Y. Chen, Y. Liu, Y. Yao, S. Zhang,

ARTICLE

Z. Gu, Q. Jin, J. Ji, X. He, K. Wang, Org. Biomol. Chem. 2017, 15, 3232–

3238.

[10] G. A. Russell, C. DeBoer, K. M. Desmond, J. Am. Chem. Soc. 1963, 85,

365–366.

[11] a) A. Podgoršek, S. Stavber, M. Zupan, J. Iskra, Tetrahedron Lett. 2006,

47, 1097–1099; b) C. H. M. Amijs, G. P. M. van Klink, G. van Koten,

Green Chem. 2003, 5, 470–474.

[12] a) M. Giroud, J. Ivkovic, M. Martignoni, M. Fleuti, N. Trapp, W. Haap, A.

Kuglstatter, J. Benz, B. Kuhn, T. Schirmeister, F. Diederich,

ChemMedChem 2017, 12, 257–270; b) Y. Wang, H. Fan, K.

Balakrishnan, Z. Lin, S. Cao, W. Chen, Y. Fan, Q. A. Guthrie, H. Sun, K.

A. Teske, V. Gandhi, L. A. Arnold, X. Peng, Eur. J. Med. Chem. 2017,

133, 197–207; c) T. V Shishkanova, M. Havlík, M. Dendisová, P. Matějka,

V. Král, Chem. Commun. 2016, 52, 11991–11994.

[13] Photochemical benzylic bromination in flow using molecular bromine: Y.

Manabe, Y. Kitawaki, M. Nagasaki, K. Fukase, H. Matsubara, Y. Hino, T.

Fukuyama, I. Ryu, Chem. Eur. J. 2014, 20, 12750–12753.

[14] Photochemical benzylic bromination in flow using NBS: a) D. Cantillo, O.

de Frutos, J. A. Rincon, C. Mateos, C. Oliver Kappe, J. Org. Chem. 2014,

79, 223–229; b) D. Šterk, M. Jukič, Z. Časar, Org. Process Res. Dev.

2013, 17, 145–151.

[15] Reviews highlighting the applicability of flow chemistry to industrial

processes: a) B. Gutmann, D. Cantillo, C. O. Kappe, Angew. Chem. 2015,

127, 6788–6832; B. Gutmann, D. Cantillo, C. O. Kappe, Angew. Chem.

Int. Ed. 2015, 54, 6688–6728; b) R. Porta, M. Benaglia, A. Puglisi, Org.

Process Res. Dev. 2016, 20, 2–25; c) D. L. Hughes, Org. Process Res.

Dev. 2018, 22, 13–20.

[16] Energy efficiency in relation to the Twelve Principles of Green Chemistry:

P. Anastas, N. Eghbali, Chem. Soc. Rev. 2010, 39, 301–312.

[17] Synthesis of this compound by benzylic bromination: a) A. Quattropani,

W. H. B. Sauer, S. Crosignani, J. Dorbais, P. Gerber, J. Gonzalez, D.

Marin, M. Muzerelle, F. Beltran, A. Nichols, K. Georgi, M. Schneider, P-

A. Vitte, V. Eligert, L. Novo-Perez, J. Hantson, S. Nock, S. Carboni, A. L.

S. de Souza, J-F. Arrighi, U. Boschert, A. Bombrun, ChemMedChem

2015, 10, 688–714; b) G. M. Ksander, E. Meredith, L. G. Monovich, J.

Papillion, F. Firooznia, Q-Y Hu, Condensed Imidazolo Derivatives for the

Inhibition of Aldosterone Synthase and Aromatase, 2007, WO

2007/024945 A1; c) Y. Ding, R. K. Thalji, J. P. Marino, Novel SHE

Inhibitors and their Use, 2009, WO 2009/049165 A1; d) J. E. Hunter, W.

C. Lo, G. B. Watson, A. Patny, G. D. Gustafson, D. Pernich, W. K.

Brewster, D. L. Camper, B. Lorsbach, M. R. Loso, T. C. Sparks, H. Joshi,

A. Mandaleswaran, R. Sanam, R. Gundla, P. S. Iyer, Pesticidal

Compositions and Process Related Thereto, 2012, WO 2012/177813 A1;

e) W. C. Lo, J. E. Hunter, G. B. Watson, A. Patny, P. S. Iyer, J. Boruwa,

Pesticidal Compositions and Process Related Thereto, 2014, US

2014/0171314 A1.

[18] B. de B. Darwent in National Standard Reference Data Series, No. 31,

National Bureau Of Standards, Washington, DC, 1970.

[19] DFT calculations (UB3LYP/6-311G++(d,p)) found the barrier for

hydrogen atom abstraction from the benzylic position of 1 and 2 to be

+6.0 kcal mol−1 and +9.1 kcal mol−1, respectively. Although there is a

significant difference between the two, both barriers are sufficiently low

to allow the process to occur in a facile manner at room temperature,

hence both are likely to be indistinguishable. See supporting information

for details.

[20] W. Offermann, F. Vögtle, Synthesis 1977, 272–273.

[21] See supporting information (Figure S25) for more detailed reaction profile

comparing different LED wavelengths.

[22] See supporting information for details of control reactions demonstrating:

no reaction in the dark, or in the presence of TEMPO, and reaction

ceasing upon removal of light source.

[23] P. Goldfinger, P. A. Gosselain, R. H. Martin, Nature 1951, 168, 30–32.

[24] See the supporting information (Figure S24) for time courses of benzylic

bromination of the model substrate with N-bromosuccinimide at 405 nm

from 30 to 400mA.

[25] Examples highlighting improved reaction performance in flow,

particularly using a microreactor: a) E. E. Coyle, M. Oelgemöller,

Photochem. Photobiol. Sci. 2008, 7, 1313; b) K. Loubière, M.

Oelgemöller, T. Aillet, O. Dechy-Cabaret, L. Prat, Chem. Eng. Process.

2016, 104, 120–132; c) L. D. Elliott, J. P. Knowles, P. J. Koovits, K. G.

Maskill, M. J. Ralph, G. Lejeune, L. J. Edwards, R. I. Robinson, I. R.

Clemens, B. Cox, D. D. Pascoe, G. Koch, M. Erlebe, M. B. Berry, K. I.

Booker-Milburn, Chem. Eur. J. 2014, 20, 15226–15232; d) K. Pimparkar,

B. Yen, J. R. Goodell, V. I. Martin, W.-H. Lee, J. A. Porco, A. B. Beeler,

K. F. Jensen, J. Flow Chem. 2011, 1, 53–55.

[26] See the supporting information for experimental characterization of the

flow reactor used in this study.

[27] See the supporting information for details of the method used (General

method A).

[28] See the supporting information for details of initial substrate screening in

batch.

[29] Synthesis and characterization data for 2-bromo-1-(4-

methylphenyl)ethanone: a) T. Ganesh, C.H. Kumar, G. L. D.

Krupadanam, Synth. Commun. 1999, 29, 2069-2078; b) B. Rammurthy,

P. Swamy, M. Naresh, C. Durgaiah, G. K. Sai, N. Narender, New J. Chem.

2017, 41, 3710-3714.

[30] See the supporting information for a temperature profile of each reaction.

[31] See the supporting information for an engineering line diagram (ELD) of

the flow reaction setup, denoting the placement of all three temperature

probes.

[32] See the supporting information (Figure S41) for details of light intensity

and flow rate optimization for the bromination of 13.

[33] See the supporting information (Figure S54 and S56) for HPLC reaction

profiles throughout both reactions.

ARTICLE

ARTICLE And then there was light: The significance of defining light power requirements within photochemical reactions has been exemplified using the Wohl-Ziegler reaction in batch and flow.

Holly E. Bonfield, Jason D. Williams, Wei Xiang Ooi, Stuart G. Leach, William J. Kerr and Lee J. Edwards*

Page No. – Page No.

A Detailed Study of Irradiation Requirements, Towards an Efficient Photochemical Wohl-Ziegler Protocol in Flow